chi meisub.chimei.org.tw/55399/images/pdf/04_announcement/2017/ap-an-106-07.pdf · rish k. pai, md,...

A Practical Approach to the Evaluation of GastrointestinalTract Carcinomas for Lynch Syndrome

Rish K. Pai, MD, PhD* and Reetesh K. Pai, MDw

Abstract: Lynch syndrome accounts for roughly 1 of every 35patients with colorectal carcinoma, making it the most commonhereditary form of colorectal carcinoma. Identifying patients atrisk for Lynch syndrome is essential, as these patients can de-velop additional Lynch syndrome–related tumors, and patientsand their relatives benefit from genetic counseling. The hallmarkof Lynch syndrome–associated neoplasms is DNA mismatchrepair protein deficiency. In most instances, the pathologist isthe first to identify patients at risk for Lynch syndrome andis tasked with communicating these results to treating cliniciansand genetics counselors. This review will attempt to provide thetools for pathologists to identify patients at risk for Lynchsyndrome through evaluation of tumors of the gastrointestinaltract and provide up-to-date knowledge on evaluating mismatchrepair protein deficiency in tumors of the gastrointestinaltract.

Key Words: Lynch syndrome, colorectal carcinoma, MSI,BRAF, mismatch repair protein, immunohistochemistry,HNPCC, Lynch-like syndrome

(Am J Surg Pathol 2016;40:e17–e34)

An estimated 130,000 individuals are diagnosed withcolorectal carcinoma each year, and approximately

50,000 will die from this disease, making colorectal car-cinoma the third leading cause of cancer-related death inthe United States.1 Lynch syndrome is defined by thepresence of a deleterious germline alteration in either aDNA mismatch repair (MMR) gene (MLH1, PMS2,MSH2, and MSH6) or deletions in EPCAM and is anautosomal dominant condition that predisposes to thedevelopment of tumors, most notably carcinomas of thecolon, rectum, and endometrium. Roughly, 1 of every 35patients with colorectal carcinoma has Lynch syndrome,making Lynch syndrome the most common hereditaryform of colorectal carcinoma. Identifying patients with

Lynch syndrome is essential as these patients and theirfamily members may be at risk for developing Lynchsyndrome–related tumors. Patients with Lynch syndromeand their affected relatives also benefit from close sur-veillance, which facilitates early cancer detection and in-tervention, which decreases disease-specific mortality.

The hallmark of Lynch syndrome–associated neo-plasms is MMR protein deficiency, manifested by eitherhigh-level microsatellite instability (MSI) identified bypolymerase chain reaction (PCR) techniques or by dem-onstration of loss of MMR protein expression within aneoplasm by immunohistochemistry. There is a continu-ing requirement for pathologists to play an increasing rolein the identification of patients at risk for Lynch syn-drome. In most instances, the pathologist is the first toidentify these patients and is tasked with communicatingthese results to treating clinicians and genetics counselors.However, in some cases, pathologists may not feel fullyproficient in explaining the implications of these findings.In addition, the terminology and definitions for con-ditions associated with MMR protein deficiency haveundergone significant refinement over the last few years.In this review, particular emphasis will be given toevolving concepts in Lynch syndrome diagnostics(Table 1), including the growing clinical problem of in-dividuals with Lynch-like syndrome, constitutionalMMR protein deficiency, and the importance of MMRprotein expression patterns in tumors in informing strat-egies for germline MMR and EPCAM gene–sequencingefforts. This review will also attempt to provide the toolsfor pathologists to identify patients at risk for Lynchsyndrome and to develop optimal screening strategies andalgorithms for detecting patients at risk for Lynch syn-drome in the setting of colorectal carcinoma and othertumors of the gastrointestinal tract.

SELECTING A SCREENING STRATEGY FORLYNCH SYNDROME: FROM MORPHOLOGY TO

UNIVERSAL SCREENINGThe optimal strategy for screening patients with

colorectal carcinoma for Lynch syndrome is a subject ofcontinued debate in the literature. Algorithms based solelyon family history are time-consuming and impractical formost physicians. The histopathologic features of color-ectal carcinomas with MMR protein deficiency have beenwell described and include increased tumor-infiltrating

From the *Department of Laboratory Medicine and Pathology, MayoClinic, Scottsdale, AZ; and wDepartment of Pathology, University ofPittsburgh Medical Center, Pittsburgh, PA.

Conflicts of Interest and Source of Funding: The authors have disclosedthat they have no significant relationships with, or financial interestin, any commercial companies pertaining to this article.

Correspondence: Reetesh K. Pai, MD, Department of Pathology, Uni-versity of Pittsburgh, Presbyterian Hospital, 200 Lothrop Street,Room A-610, Pittsburgh, PA 15213 (e-mail: [email protected]).

Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.

SPECIAL ARTICLE

Am J Surg Pathol ! Volume 40, Number 4, April 2016 www.ajsp.com | e17


Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

mailto:[email protected]

lymphocytes, a Crohn’s-like peritumoral lymphocyticreaction, mucinous/signet ring cell differentiation, andmedullary differentiation2–6 (Fig. 1). The first screeningguidelines to include a histology component are the re-vised Bethesda guidelines (Table 2).2 In a recent pooled-data analysis of 4 large cohorts, Moreira et al8 comparedvarious strategies to identify patients with Lynch syn-drome. In their study, Moreira and colleagues found thatthe revised Bethesda guidelines had a sensitivity of 87.8%for the detection of Lynch syndrome–associated colorectalcarcinomas. Moreira and colleagues also analyzed 2 ad-ditional screening approaches: (1) screening all patientswith colorectal carcinoma diagnosed in patients youngerthan 70 years based on recommendations from a 2008workshop in Jerusalem7 and (2) a selective approach ofscreening all patients with colorectal carcinoma diagnosedin patients younger than 70 years and in those above 70years fulfilling the revised Bethesda guidelines. The Jer-usalem recommendations had a sensitivity of only 85.4%,and the selective approach had a sensitivity of 95.1% in thedetection of patients with Lynch syndrome. Finally, his-tology-based models, such as MsPATH and PREDICT,have been proposed to predict MMR protein deficiency incolorectal carcinoma.9,10 Although the MsPath andPREDICT models are an improvement over the revisedBethesda guidelines, they were not specifically developedto detect Lynch syndrome–associated colorectal carcino-mas. Both place heavy emphasis on tumor location, and a

recent study demonstrated that a substantial percentage ofLynch syndrome–associated colorectal carcinomas are lo-cated in the distal colon or rectum.11 Thus, although his-tology can be helpful in identifying colorectal carcinomaswith MMR protein deficiency, models that rely solely onhistology may be less sensitive in detecting Lynchsyndrome–associated colorectal carcinoma.

Major professional medical organizations have is-sued Lynch syndrome–screening guidelines for patientswith colorectal carcinoma (Table 3). Universal screeningof all patients with colorectal carcinoma is recommendedby the Evaluation of Genomic Applications in Practiceand Prevention ( a working group sponsored by theCenters for Disease Control),13 the US Multi-SocietyTask Force for colorectal carcinoma,16 and the AmericanCollege of Gastroenterology.17 The National Compre-hensive Cancer Network, American Society of ClinicalOncology,14 and European Society of Medical Onco-logy15 guidelines also endorsed universal screening of allpatients with colorectal carcinoma; however, on the basisof the findings of Moreira et al,8 these organizations alsostate that an alternate strategy of screening all patientswith colorectal carcinoma under the age of 70 years andthose over the age of 70 years who meet the revised Be-thesda guidelines is acceptable.12 Although somewhatcost-intensive, modeling studies have shown that univer-sal screening is cost-effective with a favorable cost-effec-tiveness ratio per life-year saved.18

TABLE 1. Terminology and Definitions

Terminology Definition

Lynch syndrome (MIM No. 120435) Presence of deleterious monoallelic germline defect in MMR genes including:1. Germline mutations in MLH1, PMS2, MSH2, and MSH62. Inactivation of MSH2 due to deletion of 3’ exons of the EPCAM (TACSTD1) gene3. Germline methylation of the MLH1 promoter (also called constitutional MLH1 inactivation)The genetic definition of Lynch syndrome encompasses Muir-Torre syndrome (MIMNo. 158320) definedby the association of sebaceous gland tumors with a Lynch syndrome–associated internal malignancy

HNPCC Clinical term for patients with carcinoma that fulfill the following Amsterdam I or II clinical criteria:1. Three or more family members (one of whom is a first-degree relative of the other 2) with colorectal

carcinoma (Amsterdam I) or HNPCC-related cancers (Amsterdam II), which include colorectal,endometrial, stomach, small intestinal, hepatobiliary, renal pelvis, or ureteral cancers

2. At least 2 successive affected generations3. At least one of the HNPCC-related cancers diagnosed before the age of 504. Familial adenomatous polyposis (FAP, MIM No. 175100) should be excluded

Lynch-like syndrome Patients who have tumors demonstrating deficient MMR protein expression but have no deleteriousgermline alteration in MMR genes or EPCAM, and if the tumor is MLH1-deficient no evidence of aBRAF V600E mutation orMLH1 promoter hypermethylation within the tumor. Potential explanationsfor Lynch-like syndrome include:

1. Biallelic somatic (tumor) mutations in the affected MMR gene2. Failure to detect germline MMR gene alteration using currently available testing methods3. Incorrect interpretation of MMR immunohistochemistry results4. Other germline gene defects, such as biallelic MUTYH mutation

Familial colorectal carcinoma (type X) Clinical term for patients with carcinoma that fulfill the Amsterdam I clinical criteria but have tumorsthat do not harbor MMR protein deficiency and whose germline lacks MMR gene or EPCAMalterations. The genetic basis for this condition is unknown

Constitutional mismatch repair deficiency(MIM No. 276300)

Presence of biallelic germline mutations in MMR genes. This disease is very rare with a propensity topresent with malignancy early in life

Sporadic MLH1 deficiency ingastrointestinal tract cancer

Somatic silencing of MLH1 protein expression. This is most commonly caused by epigenetic MLH1promoter hypermethylation. In the colon/rectum, a BRAF V600E mutation indicates sporadic MMRdeficiency but is only present in 50% to 70% of tumors withMLH1 promoter hypermethylation. BRAFV600E mutation cannot be used as a surrogate for MLH1 promoter hypermethylation in extracoloniccarcinomas

Pai and Pai Am J Surg Pathol ! Volume 40, Number 4, April 2016

e18 | www.ajsp.com Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.


Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

SCREENING ALGORITHM FOR LYNCHSYNDROME IN COLORECTAL CARCINOMA

Figure 2 details a screening algorithm for Lynchsyndrome in colorectal carcinoma using MMR proteinimmunohistochemistry as the initial screening method.The details of each individual test are described below.

MMR Protein Immunohistochemistry and itsPitfalls

Early reports initially suggested that MMR proteinimmunohistochemistry was inferior to MSI PCR as ascreening tool for Lynch syndrome. However, these earlystudies primarily assessed MLH1 and MSH2 immuno-histochemistry without using antibodies directed againstMSH6 and PMS2. More recent literature has demon-strated that the use of all 4 antibodies (MLH1, MSH2,MSH6, and PMS2) has a high sensitivity (ranging from93% to 100%) for detecting high-level MSI and for pre-dicting MMR gene mutation.19–21 Knowledge of the basicbiology of MMR proteins helps to better understand the

immunohistochemical staining patterns often observed intumors with MMR protein deficiency. In their functionalstate within a cell, MLH1 dimerizes with PMS2, andMSH2 dimerizes with MSH6. MLH1 and MSH2 are theobligatory partners for their respective heterodimers. Ingeneral, mutations in MLH1 and MSH2 result in pro-teolytic degradation of their binding partners, PMS2 andMSH6, respectively. In contrast, mutations in MSH6 andPMS2 do not result in proteolytic degradation of theirobligatory partners, as other MMR proteins have over-lapping function with those of PMS2 and MSH6. Thus, acancer from a patient with a germline mutation in MSH2typically exhibits loss of expression of both MSH2 andMSH6, whereas a cancer from a patient with a germlineMSH6 mutation exhibits isolated loss of MSH6 ex-pression.

Given the biology of MMR proteins, a number ofstudies have analyzed an initial screening approach usinga 2-antibody panel including only PMS2 and MSH6.These studies have demonstrated that this 2-antibody

FIGURE 1. Histologic features associated with MMR protein deficiency in colorectal carcinoma proposed by the revised Bethesdaguidelines include tumor-infiltrating lymphocytes (A) that can be seen in carcinomas with either a glandular or medullary growthpattern (A), Crohn’s-like peritumoral lymphocytic reaction (B), mucinous differentiation (C), and signet ring cell differentiation(D).

Am J Surg Pathol ! Volume 40, Number 4, April 2016 Lynch Syndrome in the Gastrointestinal Tract

Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved. www.ajsp.com | e19


Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

approach is as effective as using the 4-antibody panel fordetecting patients with colorectal carcinoma at risk forLynch syndrome.22–24 MMR protein immunohistochemistryusing either a 2-antibody or a 4-antibody screening approachis acceptable. Most institutions still use the 4-antibodyMMR protein immunohistochemistry panel for Lynch syn-drome screening in colorectal carcinoma, as PMS2 andMSH6 immunohistochemistry can be technically challengingto optimize and can at times be difficult to interpret.

There are 3 common patterns of MMR protein im-munohistochemical expression: diffuse preserved (intact)

expression; complete loss of expression; and heterogenousexpression (Fig. 3). When reporting MMR protein im-munohistochemistry results, it is important to avoid termssuch as “positive” and “negative,” as they are potentiallyconfusing. Instead, terminology such as preserved (or in-tact) or loss of expression is preferred. Most colorectal

TABLE 2. The Revised Bethesda Guidelines for TestingColorectal Carcinomas for Mismatch Repair ProteinDeficiency*

Tumor from individuals should be tested for mismatch repair proteindeficiency in the following situations:1. Colorectal carcinoma diagnosed in a patient who is below 50 y ofage

2. Presence of synchronous, metachronous colorectal, or other Lynchsyndrome–associated tumorsw, regardless of age

3. Colorectal carcinoma with MSI-H histologyz diagnosed in a patientwho is below 60 y of agey

4. Colorectal carcinoma diagnosed in one or more first-degreerelatives with a Lynch syndrome–related tumor, with one of thecancers being diagnosed under age 50 y

5. Colorectal carcinoma diagnosed in 2 or more first-degree or second-degree relatives with Lynch syndrome–related tumors, regardless ofage

*Proposed by Umar et al.2

wLynch syndrome–related tumors include colorectal, endometrial, stomach,ovarian, pancreas, ureter and renal pelvis, biliary tract, brain tumors (usuallyglioblastoma), sebaceous tumors of the skin, keratoacanthomas, and carcinoma ofthe small bowel.zPresence of tumor-infiltrating lymphocytes, Crohn’s-like lymphocytic re-

action, mucinous/signet ring cell differentiation, or medullary growth pattern.yThere was no consensus among the Workshop participants on whether to

include the age criteria in guideline 3 above; participants voted to keep below 60years of age in the guidelines.

Adaptations are themselves works protected by copyright. So in order topublish this adaptation, authorization must be obtained both from the owner ofthe copyright in the original work and from the owner of copyright in the trans-lation or adaptation.

TABLE 3. Summary of Lynch Syndrome ScreeningRecommendations From Professional Organizations

Screening Strategy forPatients With ColorectalCarcinoma

Sensitivity forDetecting LynchSyndrome (%)*

Major ProfessionalOrganizations

Endorsing Strategy

Universal screening 100 NationalComprehensiveCancer Network12

Evaluation of GenomicApplications inPractice andPrevention13

American Society ofClinical Oncology14

European Society forMedical Oncology15

US Multi-Society TaskForce on ColorectalCancer16

American College ofGastroenterology(ACG)17

Screening of all patientsbelow 70 y of age and inpatients above 70 yfulfilling revisedBethesda guidelines

95.1 NationalComprehensiveCancer Network12

American Society ofClinical Oncology14

European Society forMedical Oncology15

Revised Bethesdaguidelines

87.8 None

*Sensitivities based on Moreira et al.8

FIGURE 2. A screening algorithm for Lynch syndrome in colorectal carcinoma using MMR protein immunohistochemistry as theinitial screening test. A, With a universal screening approach, every patient with colorectal carcinoma will undergo MMR proteinimmunohistochemical analysis using either a 4-antibody (MLH1, PMS2, MSH2, and MSH6) or a 2-antibody (PMS2 and MSH6)approach. Most tumors will demonstrate preserved expression of all tested MMR proteins and will require no further testing.Patients with tumors demonstrating concurrent loss of MSH2 and MSH6, isolated loss of MSH6, or isolated loss of PMS2expression should be referred for genetic counseling and germline MMR gene and/or EPCAM mutation testing. Patients withtumors demonstrating loss of MLH1 and PMS2 expression should have reflex BRAF mutation testing and, if BRAF wild-type, MLH1promoter hypermethylation testing of the tumor with subsequent referral to genetic counseling if negative for MLH1 promoterhypermethylation. In the setting of positive MLH1 promoter hypermethylation studies, some authors argue for evaluation forconstitutional MLH1 epimutation if the patient is below 60 years of age and/or if there is a strong personal or family history ofLynch syndrome–associated malignancy. B, If a deleterious germline MMR gene or EPCAM mutation is identified, the patient hasLynch syndrome. If no deleterious germline mutation in an MMR gene or EPCAM is identified, the patient has “Lynch-like”syndrome. Rereview of the original MMR protein immunohistochemistry should be performed to confirm loss of MMR proteinexpression within the patient’s tumor. Somatic MMR gene mutation may be of benefit. If biallelic somatic MMR gene mutation isidentified, the patient does not have Lynch syndrome and can likely be screened as for sporadic colorectal carcinoma. If there is noevidence of biallelic somatic mutation, there is uncertainty as to whether the patient has Lynch syndrome. Most patients withLynch-like syndrome lacking biallelic somatic mutation are counseled to follow a screening protocol as if they have confirmedLynch syndrome given the uncertainty in the diagnosis.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

carcinomas will demonstrate diffuse, strong nuclear ex-pression of MMR proteins by immunohistochemistry andcan be labeled as having preserved expression for MMRproteins. Loss of expression for MMR proteins is definedas complete absence of nuclear staining within tumor cellswith concurrent positive labeling in internal non-neoplastic

tissues. Care must be taken to not misinterpret nuclearstaining within tumor-infiltrating lymphocytes as tumorcells exhibiting nuclear expression within tumor cells.

There are many pitfalls in MMR protein im-munohistochemistry interpretation, particularly withregard to aberrant staining patterns and variability in

A

B

Patientwith

ColorectalCarcinoma

MMR IHC

PreservedExpression of

all 4 MMRproteins

No furthertesting required

Loss ofMLH1

expression

BRAFMutationTesting

BRAF V600EMutationPositive

Wild-typeBRAF

MLH1 PromoterHypermethylationAnalysis of Tumor

Negative for MLH1Promoter Hypermethylation

in Tumor

GermlineMMR Gene and/orEPCAM Mutation

Testing

Loss of MSH2/MSH6,Isolated loss of MSH6, or

Isolated loss of PMS2

Referral to GeneticCounseling

Positive for MLH1Promoter Hypermethylation

in Tumor

Referral to GeneticCounseling

GermlineMMR Gene and/orEPCAM Mutation

Testing

Deleterious MMRgene or EPCAM

germline mutationidentified

Lynch SyndromeConfirmed

IntensiveLifelong

Screening

NO Deleterious MMRgene or EPCAM

germline mutationidentified

“Lynch-Like”Syndrome

Confirm OriginalMMR IHC Results

Somatic MMR GeneMutation Testing of Tumor

Biallelic SomaticMutation Absent

Biallelic SomaticMutation Present

Screening as forsporadic carcinoma




staining intensity. Variation in staining patterns has re-sulted in a less than perfect interobserver agreement in theinterpretation of MMR protein immunohistochemicalstains.19,25 Aberrant staining patterns include punctate/speckled nuclear staining, nucleolar staining,26 and nu-clear membrane staining (Figs. 3, 4). In most cases, thesestaining patterns are a result of technical issues andshould not be interpreted as preserved (intact) expressionof MMR proteins. Weak staining can also be a diagnosticchallenge. Comparison with the internal control is often

helpful to point to technical issues. However, in somecases, weak MLH1 staining may actually be a result of anMLH1 gene mutation.19

The most common heterogenous pattern is the ab-sence of MMR protein expression in scattered tumor cellsin a background of strong, diffuse expression within theremainder of the tumor (intratumoral heterogeneity).Tissue hypoxia and fixation issues may account for thisparticular pattern of expression.27,28 Neoadjuvant che-moradiation may also reduce the expression of MMR

FIGURE 3. Typical patterns of expression of MMR proteins by immunohistochemistry in colorectal carcinoma. Most colorectalcarcinomas will demonstrate diffuse, strong nuclear expression of MMR proteins by immunohistochemistry and can be labeled ashaving preserved expression for MMR proteins (A, MSH2 immunohistochemistry). Loss of expression for MMR proteins is definedas complete absence of nuclear staining within tumor cells with concurrent positive labeling in internal non-neoplastic tissues.Care must be taken to not misinterpret nuclear staining within tumor-infiltrating lymphocytes as tumor cells exhibiting nuclearexpression within tumor cells. This is particularly important for colorectal carcinomas with prominent tumor-infiltrating lym-phocytes, such as this case of medullary carcinoma that demonstrates loss of PMS2 expression within tumor cells with concurrentstrong nuclear expression within tumor-infiltrating lymphocytes (B). Patchy, heterogenous expression of MMR proteins withincolorectal carcinoma can be seen by immunohistochemistry. This is particularly common with MSH6 immunohistochemistry (C).Heterogenous staining should not be interpreted as evidence of loss of MMR protein expression. If >10% of tumor nucleidemonstrate convincing and strong nuclear expression, the tumor should be labeled as demonstrating preserved MMR proteinexpression. In some instances, nuclear staining that may be weak is observed in <10% of tumor nuclei (D, MSH6). In such cases,the MMR protein immunohistochemistry staining should be interpreted as equivocal. Microsatellite instability polymerase chainreaction analysis of the tumor and/or genetic counseling may be of benefit for patients with equivocal MMR expression withintheir tumor.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

proteins. Bao et al29 compared immunohistochemicalexpression of MMR proteins in pretreatment biopsies andposttreatment resections in 51 rectal carcinomas. DecreasedMSH6 expression was seen in 20% of posttreatment speci-mens compared with the pretreatment biopsies.29 In somecases, up to 90% of tumor cell nuclei had loss of MSH6expression despite demonstrating intact expression in thepretreatment biopsy. More recently, Vilkin et al30 demon-strated decreased expression of PMS2 in 30% of posttreat-ment rectal carcinomas. Decreased expression of MLH1,MSH6, and MSH2 was also seen in a small proportion ofcases. Given these findings, most pathologists with experi-ence in interpreting MMR protein immunohistochemistry

regard convincingly strong expression of MMR proteinswithin >10% of the tumor as indicative of preserved (intact)expression.

Zonal heterogeneity is another pattern occasionallyseen in colorectal carcinomas31 and more recently de-scribed in endometrial carcinomas.32 In this pattern, largeareas of tumor demonstrate abrupt loss of expression of aparticular MMR protein (Fig. 4). Abrupt loss of MLH1and PMS2 expression within large areas of a tumor istypically the result of hypermethylation of the MLH1promoter within different parts of the tumor.31 This pat-tern of abnormal MLH1 and PMS2 expression is notassociated with Lynch syndrome. Another frequently

FIGURE 4. Unusual patterns of immunohistochemical expression of MMR proteins in colorectal carcinoma. Punctate/specklednuclear staining can be seen, particularly with MLH1 immunohistochemistry (A). This pattern of staining should not be in-terpreted as preserved MLH1 expression and is likely a technical issue with the MLH1 staining protocol. This pattern of punctate/speckled nuclear staining with MLH1 is often associated with loss of PMS2 expression within the tumor and with MLH1 promoterhypermethylation and BRAF V600E mutation. Prominent and diffuse nucleolar MSH6 expression may also be seen (B) and shouldnot be interpreted as evidence of preserved MSH6 expression. In cases of MSH6 nucleolar expression, microsatellite instabilitypolymerase chain reaction analysis of the tumor and genetic counseling should be considered. Rare colorectal carcinomas maydemonstrate distinct staining of the nuclear membrane of tumor cell nuclei without diffuse nuclear staining (C, courtesy ofDr Thomas Plesec, Cleveland Clinic). Nuclear membrane staining alone should not be interpreted as preserved MMR proteinexpression and is likely a technical issue with the staining protocol.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

encountered scenario is loss of MSH6 expression in tumorsthat also demonstrate concurrent loss of MLH1 and/orPMS2 expression. The loss of MSH6 expression in suchcases can be abrupt and seen in only parts of the tumorand, in most cases, is due to a secondary mutation in acoding mononucleotide microsatellite within the MSH6gene as a result of MMR protein deficiency due to loss ofMLH1 and/or PMS2 expression within the tumor.33

Practically, one should be cautious in issuing a reportwhen an unusual pattern of MMR protein expression isseen. In cases with an aberrant pattern of expression (cy-toplasmic, nuclear membrane, or nucleolar expression),repeating the immunohistochemistry, preferably on an-other tissue block, may be helpful. If a definitive result isstill not obtained, performing MSI testing by PCR may beof use. Most cases with intratumoral heterogeneity can besafely interpreted as having preserved (intact) expression.When only rare tumor cells (<10% of tumor cell nuclei)demonstrate nuclear expression for a given MMR protein,MSI testing by PCR should also be considered. The rarecases with marked zonal heterogeneity or loss of an un-usual combination of MMR proteins (eg, MLH1, PMS2,and MSH6) often need additional molecular testing.Testing for BRAF mutation and/or MLH1 promoter hy-permethylation studies can help to exclude Lynch syn-drome in these scenarios, although recently a case of Lynchsyndrome with a deleterious MSH6 mutation and MLH1promoter hypermethylation was described.34 Close com-munication with the patient’s treating clinicians and ge-netic counselor can also help guide the subsequent steps inthese unusual situations.

The pattern of abnormal MMR protein expressionwithin a patient’s tumor helps to direct germline mutationtesting efforts in a patient. However, current guidelines ongermline mutation testing are not entirely clear for thescenario of isolated loss of PMS2 immunohistochemicalexpression. Isolated loss of PMS2 expression accounts forapproximately 4% of colorectal carcinomas with abnormalMMR protein expression. Recent data indicate that ap-proximately 25% of patients with tumors demonstratingisolated loss of PMS2 immunohistochemical expression willharbor germline MLH1 mutations.35 A subset of germlineMLH1 mutations results in functionally inactive MLH1protein, which is antigenically intact and will be detected bycommonly used anti-MLH1 antibody clones (Fig. 5).Most such MLH1 mutations are nontruncating point mu-tations or mutations in noncoding intervening sequencesthat result in decreased MLH1 protein stability and/orquantity, compromised stability of the MLH1-PMS2complex, and subsequent PMS2 degradation. For suchpatients, falsely preserved MLH1 protein expression will beobserved in a patient’s tumor; however, diminished or weakMLH1 staining intensity may be a clue that an underlyinggermline MLH1 mutation is the cause of the abnormalMMR protein expression. Thus, patients with colorectalcarcinoma demonstrating isolated loss of PMS2 by im-munohistochemistry should undergo germline MLH1analysis if no mutations are detected through germlinePMS2 testing.

MSI PCRIn 1997, the National Cancer Institute recom-

mended a panel of 5 microsatellites, referred to as theBethesda panel, which included 2 mononucleotide repeats(BAT25 and BAT26) and 3 dinucleotide repeats (D2S123,D5S346, and D17S250). The Bethesda panel is still one ofthe most widely used panels of microsatellites. However,many investigators have found mononucleotide markersto be more sensitive than dinucleotide markers for thedetection of MSI, and commercially available panels ofquasimonomorphic mononucleotide markers are beingincreasingly used.36 On the basis of established criteria forthe Bethesda panel of microsatellites, if no markers showinstability, the tumor is classified as microsatellite stable.If Z30% of the markers show instability by PCR, the tu-mor is classified as high-level MSI (MSI-H). Occasionally,when using the Bethesda panel, colorectal carcinoma maydemonstrate instability in <30% of microsatellite markers.In such instances, the tumor can be designated as low-levelMSI (MSI-L).37 Most MSI-L colorectal carcinoma willdemonstrate preserved/intact MMR protein expression byimmunohistochemistry.38 However, mutations in MSH6primarily induce instability in mononucleotide and not di-nucleotide repeats. A small subset of patients with tumorsharboring MSI-L with the Bethesda panel of microsatellitemarkers may demonstrate loss of MSH6 expression withintheir tumor by immunohistochemistry and harbor a germ-line MSH6 mutation.38 Panels of mononucleotide markershave been shown to be more sensitive in detecting MSI-Hcaused by MSH6 mutations.39

Detection of MSI in colorectal carcinoma by usingnext-generation sequencing methodology has also beenrecently described.40 With increasing use of next-gen-eration sequencing for detection of molecular mutationsin colorectal carcinoma to predict response to certaintherapy, next-generation sequencing methods to detectMSI may replace the current method of detecting MSI byconventional PCR techniques.40

Selecting a Screening Test: MMR ProteinImmunohistochemistry Versus MSI PCR

Both MMR protein immunohistochemistry and MSIPCR are equally valid initial screening tests for Lynchsyndrome. The decision of which screening test to use pri-marily depends on the availability of resources and ex-pertise. However, MMR protein immunohistochemistry hasseveral advantages over MSI PCR. As MSI PCR typicallyrequires normal tissue for interpretation, it has limitedutility in biopsies in which normal tissue is not sampled.More importantly, MMR protein immunohistochemistryhelps direct germline mutation testing efforts, whereas MSIPCR does not. In addition, MMR protein immunohisto-chemistry allows for the distinction between sporadic andLynch syndrome–associated neoplasms, as loss of ex-pression of proteins other than MLH1 raises concern forLynch syndrome. Figure 2 details a screening algorithmusing MMR protein immunohistochemistry as the screeningtest. If MSI PCR is used as the initial screening test, theidentification of MSI-H within a tumor typically requires




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

further evaluation of the tumor by MMR protein im-munohistochemistry to determine which MMR proteins aredeficient within the patient’s tumor.

Sporadic MLH1 Deficiency in ColorectalCarcinoma From Promoter Hypermethylation

Sporadic MMR protein deficiency accounts forapproximately 10% to 15% of colorectal carcinoma andfrequently arises due to hypermethylation of CpG islandsin the MLH1 promoter. The majority of colorectal car-cinomas with sporadic MMR protein deficiency occur inindividuals above 60 years of age with an approximately 2to 3:1 female predominance. Acquired methylation ofboth copies of the MLH1 promoter leads to MLH1 genesilencing and concurrent loss of MLH1 and PMS2 pro-tein expression. The BRAF V600E mutation is seen inapproximately 50% to 70% of colorectal carcinomas with

MLH1 promoter hypermethylation and is very seldomidentified in tumors in patients with Lynch syn-drome.11,41–44 Thus, screening tumors with loss of MLH1and PMS2 protein expression for a BRAF V600E muta-tion is a simple way to exclude Lynch syndrome in pa-tients (Fig. 2).43 If the tumor harbors the BRAF V600Emutation, further counseling and testing for MMR genemutation is not necessary, provided there is no strongfamily or personal history of Lynch syndrome–associatedneoplasia. Importantly, the absence of a BRAF V600Emutation does not imply Lynch syndrome. If a colorectalcarcinoma is BRAF wild-type, further testing for MLH1promoter hypermethylation can be performed and, ifpositive, helps to exclude Lynch syndrome (Fig. 2). BRAFmutation analysis is only helpful in the Lynch syndrome–screening algorithm of colorectal carcinoma and is notuseful for extracolonic cancers. Extracolonic cancers of

FIGURE 5. Some cases of colorectal carcinoma (A, hematoxylin and eosin) demonstrate weak MLH1 expression compared withinternal positive control stromal cells (B) and complete loss of PMS2 expression (C). In this case, 50% of the tumor demonstratedMLH1 protein expression. The tumor was confirmed to have high-level microsatellite instability by polymerase chain reaction.Germline mutation analysis identified an MLH1 point mutation in this patient. It is likely that the germline MLH1 point mutation inthis patient resulted in functionally inactive MLH1 protein that is antigenically intact and is detected by commonly used anti-MLH1 antibody clones.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

the gastrointestinal tract may harbor MLH1 promoterhypermethylation that is not associated with a BRAFV600E mutation.

A monoclonal antibody (VE1) targeting the BRAFV600E mutant protein has become available with variableefficacy in literature reports.45–47 Most studies have dem-onstrated that BRAF VE1 immunohistochemistry is sensi-tive for detecting colorectal carcinomas harboring a BRAFV600E mutation. However, there are reports of significantvariability in staining with the BRAF VE1 antibody, par-ticularly between new antibody lots purchased from thevendor or when new reagents were used in the assay.46 Inaddition, some colorectal carcinomas can demonstrate weakstaining with BRAF VE1 immunohistochemistry but lackthe BRAF V600E mutation on PCR. Given the variabilityof staining and the issue of weak, nonspecific staining insome BRAF wild-type colorectal carcinomas, BRAF VE1immunohistochemistry is not currently commonly used inLynch syndrome diagnostics.

Patient Follow-up of Abnormal Test ResultsImplementation of universal screening requires

close cooperation and effective communication acrossmultiple clinical disciplines. In most instances, the pathol-ogist is the first to identify patients at risk for Lynchsyndrome and is tasked with communicating these resultsto treating clinicians and genetics counselors. In 1 study,communication of the results solely to colorectal surgeonsresulted in only 32% of patients at risk for Lynch syn-drome undergoing genetic counseling.48 In contrast,communication of the results directly to genetics counselorsresulted in 71% of patients at risk for Lynch syndromeundergoing genetic counseling.48 Before implementation ofuniversal screening for Lynch syndrome in colorectal car-cinoma, pathologists are strongly encouraged to establishclose communication with surgeons, medical oncologists,gastroenterologists, and especially with genetics counselorswithin their community to develop effective strategies foridentifying patients at risk for Lynch syndrome and es-tablishing appropriate plans for patient contact and patientfollow-up. In addition, if a patient at risk for Lynch syn-drome is identified by a pathologist during universalscreening, the communication of the results to treatingclinicians and/or genetics counselors responsible for patientfollow-up should be explicitly documented in the pathologyreport.

ISSUES IN LYNCH SYNDROME DIAGNOSTICS

Definition of Lynch Syndrome and its RelatedConditions

Table 1 provides a summary of definitions for theoften confusing and changing terminology used for Lynchsyndrome and its related conditions. Lynch syndrome isdefined by the presence of a deleterious germline alterationin either a DNA MMR gene (MLH1, PMS2, MSH2, andMSH6) or deletions in EPCAM. Lynch syndrome andhereditary nonpolyposis colorectal carcinoma (HNPCC)have been historically linked; however, these 2 conditions

are not synonymous. HNPCC is a clinical term for patientswith carcinoma that fulfill Amsterdam clinical criteria thatare based solely on family history (Table 1).49 Between50% and 60% of patients with HNPCC will have tumorsthat are MMR protein deficient.50 Approximately 40% ofpatients with HNPCC do not harbor MMR protein defi-ciency within their tumor or have a germline DNA MMRgene or EPCAM alteration and have been labeled ashaving familial colorectal carcinoma type X (FCCTX)(Table 1).51,52 Unlike Lynch syndrome, patients withFCCTX do not have an increased risk for extracoloniccancers but have a 2-fold increased risk for colorectalcarcinoma over the general population.50,52 The geneticetiology of FCCTX is largely unknown, but candidategenes are currently under investigation.52–55

Lynch Syndrome Versus “Lynch-like” SyndromeAlthough loss of MMR protein expression within a

patient’s tumor raises concern for Lynch syndrome, itshould not be considered diagnostic of Lynch syndrome.Some patients have no evidence of a germline alterationinMMR genes or EPCAM despite harboring tumors withMMR protein deficiency. Lynch-like syndrome is a pro-visional term that has emerged in the literature to describethose patients who have tumors demonstrating deficientMMR protein expression but have no deleterious germ-line alteration in MMR genes or EPCAM and, if the tu-mor is MLH1 deficient, no evidence of a BRAF V600Emutation or MLH1 promoter hypermethylation withinthe tumor.56 This group of patients has also been termed“suspected Lynch syndrome” by others.57 The terms“Lynch-like syndrome” or “suspected Lynch syndrome”are probably not the best terms to describe patients whodevelop tumors with MMR protein deficiency throughmechanisms other than germline MMR gene or EPCAMalteration or MLH1 promoter hypermethylation. However,for the purposes of this discussion, the term Lynch-likesyndrome will be used to describe these patients.

Patients with Lynch-like syndrome represent a majorchallenge in Lynch syndrome screening as it is uncertainwhether such patients should undergo the same intensivelifelong screening protocol used for patients with confirmedLynch syndrome. The proportion of patients with Lynch-like syndrome among patients with tumors demonstratingMMR protein deficiency concerning for Lynch syndromevaries in the literature. A recent review by Buchanan et al58

found that up to 59% of colorectal carcinomas with MMRprotein deficiency have been identified as having Lynch-likesyndrome, with the proportion ranging from 56% to71%.21,56,57 In our experience with universal screening incolorectal carcinoma, 32% of patients with colorectal car-cinoma harboring abnormal MMR protein expression byimmunohistochemistry within their tumor raising concernfor Lynch syndrome have been negative for a germlinemutation in MMR genes or EPCAM and have been clas-sified as having Lynch-like syndrome.59

There are a number of potential reasons for Lynch-like syndrome (Table 4). First, and easiest to evaluate, ismisinterpretation of MMR protein immunohistochemistry.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

When faced with a discrepancy between the MMR proteinimmunohistochemistry results and germline testing, it isprudent to reevaluate the MMR protein immunohisto-chemical stains and possibly repeat any stains that areequivocal to confirm the presence of abnormal MMR pro-tein expression within the tumor. In 1 analysis of patientswith Lynch-like syndrome, 19% were found to be the resultof incorrect interpretation of MMR protein immuno-histochemistry.7

Some patients with Lynch-like syndrome likely stillhave a germline mutation that current testing methodsfailed to detect. An excellent example is that of the re-cently identified inversion of MSH2 exons 1 to 7, whichhave been found to explain up to 60% of previously un-explained cases of Lynch syndrome due to MSH2 germ-line mutations.60 Inversion of MSH2 exons 1 to 7 has notbeen specifically analyzed by many commercial geneticlaboratories or has only recently been added to the testingprotocol of other laboratories. In addition, there havebeen reports of somatic mosaicism in which the patientharbors a deleterious mutation in only a portion of theircells that may explain Lynch-like syndrome in rare cases.61

Finally, other germline genetic defects may explain Lynch-like syndrome. In 1 recent study, approximately 3% ofpatients with Lynch-like syndrome had biallelic MUTYHmutations.62

More recent literature reports have identified bial-lelic mutations in MMR genes in most tumors in patientswith Lynch-like syndrome.61,63–65 Thus, these patients donot have Lynch syndrome but have somatic (tumor)MMR gene mutations that explain the MMR proteindeficiency within their tumor. In 1 series, Haraldsdottirand colleagues analyzed tumors with deficient MMRprotein immunohistochemical expression from 32 pa-tients with no germlineMMR gene mutation or alterationand performed somatic MMR gene mutation analysis.Twenty-two of 32 (69%) tumors had 2 somatic mutationsin MMR genes that explained the loss of MMR proteinexpression by immunohistochemistry. Overall, approx-imately 70% of patients who lack a germline MMR genemutation or alteration and who harbor tumors demon-strating loss of MMR protein expression can be explainedby biallelic somatic MMR gene mutations in their tumor.Interestingly, on the basis of the accumulated data fromthe literature, different patterns of somatic mutations areevident based on which MMR genes are affected. ForLynch-like tumors with loss of MLH1 and PMS2 proteinexpression by immunohistochemistry, the most commonscenario is a somatic MLH1 gene mutation (deletion,frameshift, insertion, or duplication) associated with lossof heterozygosity of the other MLH1 allele. In contrast,for Lynch-like tumors with loss of MSH2 and MSH6protein expression by immunohistochemistry, the mostcommon scenario is biallelic somatic MSH2 gene muta-tion with infrequent loss of heterozygosity of MSH2. Toofew cases of patients harboring tumors with isolated lossof either MSH6 or PMS2 have been analyzed in the lit-erature to comment on patterns of somatic mutation, ifany. Somatic MMR gene mutations are likely sporadicevents; however, the possibility that these somatic biallelicmutations are secondary to other germline gene defects,including MUTYH, still remains.62

There also appear to be significant clinicopathologicdifferences between patients with Lynch-like syndromeand Lynch syndrome.59 Colorectal carcinoma in patientswith Lynch-like syndrome is even more likely to be locatedin the proximal colon. Certain patterns of abnormalMMR protein expression are also more likely to haveLynch syndrome confirmed by germline mutation anal-ysis. Most patients with colorectal carcinomas demon-strating isolated loss of MSH6 expression will have Lynchsyndrome confirmed by germline mutation analysis. Incontrast, Lynch-like syndrome is more often identified inpatients with tumors that harbor concurrent MLH1/PMS2 loss (with lack of MLH1 promoter hyper-methylation) or concurrent MSH2/MSH6 loss. Synchro-nous or metachronous Lynch syndrome–associatedcarcinomas are more frequently identified in patients withconfirmed Lynch syndrome compared with Lynch-likesyndrome. Similar to colorectal carcinomas in patientswith confirmed Lynch syndrome, carcinomas in patients

TABLE 4. Potential Explanations for Lynch-like Syndrome

ReasonProportionof Cases Explanation

Biallelic (somatic) MMRgene mutation within thetumor

B70% Biallelic somatic MMR genemutations within the tumoroccur most frequently inMLH1 and MSH2. Theseare likely sporadic events,and such patients can bescreened as if they hadsporadic colorectalcarcinoma

Incorrect interpretation ofMMR proteinimmunohistochemistry

B20% Reevaluation of MMRproteinimmunohistochemicalstains and repeating anyequivocal staining canidentify false-“positive”results

MMR gene mutationspresent but not detectedby current testingmethods

Unknown One example is the inversionof MSH2 exons 1 to 7 thathas recently been identifiedto explain some cases ofLynch syndrome. Thisinversion has not beenroutinely tested by mostlaboratories

Germline mutations inother genes, such asMUTYH

3% Germline MUTYH mutationanalysis may be consideredin patients with Lynch-likesyndrome who lack biallelicsomatic MMR genemutation

Somatic mosaicism <1% Very rare. Most next-generation sequencingmethodologies can detectmutations in <5% ofDNA, making thepossibility of undetectedmosaicism unlikely




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

with Lynch-like syndrome develop through the conven-tional adenoma pathway and not the serrated pathway.This indicates that colorectal carcinomas in patients withLynch-like syndrome may arise from conventional ad-enomas that develop biallelic somatic MMR gene muta-tions resulting in colorectal carcinoma with MMR proteindeficiency.

EPCAM Deletions in Lynch SyndromeGermline deletions affecting the EPCAM (TACSTD1)

gene can lead to epigenetic silencing of the MSH2 gene byhypermethylation and cause Lynch syndrome.66,67 EPCAMdeletions with associated hypermethylation of the MSH2promoter are thought to account for up to 20% to 25% ofLynch syndrome tumors with loss of MSH2 and MSH6.67,68

Patients with tumors harboring loss of MSH2 and MSH6expression should be tested for germline deletions in EPCAM,if no MSH2 germline mutation is identified. Immuno-histochemistry using the monoclonal antibody Ber-EP4directed at EPCAM has been proposed as helping to differ-entiate patients with germline EPCAM deletion from thosewith germline MSH2 mutation.69 In 1 analysis, patients withEPCAM deletions demonstrated loss of Ber-EP4 expressionwithin the tumor, whereas none of the patients with MSH2germline mutations displayed loss of Ber-EP4 expression.69 Asubsequent study from the same group demonstrated that lossof Ber-EP4 expression occurs in many, but not all, tumorsfrom patients with Lynch syndrome caused by EPCAMgermline deletions.70 An acquired somatic deletion of thesecond EPCAM allele within the tumor is required for loss ofBer-EP4 expression. Those tumors that do not acquire asecond somatic EPCAM deletion in addition to the germlineEPCAM deletion will demonstrate preserved expression ofBer-EP4 within their tumor.70 Thus, Ber-EP4 immuno-histochemistry is not particularly helpful in identifyingpatients with Lynch syndrome due to germline EPCAM de-letion.

Constitutional MLH1 EpimutationConstitutional epimutation refers to epigenetic hy-

permethylation at the promoter of a gene leading to si-lencing of expression from the allele within all normalsomatic tissues. Constitutional epimutation of MLH1 hashelped to explain a subset of Lynch syndrome patients. Inthese patients, the MLH1 promoter of the affected allele isfully methylated, resulting in silencing of the allele withinall or most cells within the body. The second “hit” of theMLH1 gene in these patients is often loss of heterozygosityor acquired sequence mutations of the secondMLH1 alleleresulting in complete loss of MLH1 protein expressionwithin affected cells.71 Some constitutional MLH1 epi-mutations are labile in the germline, and transmission tosuccessive generations may not occur. Thus, constitutionalMLH1 epimutation may not follow classical Mendelianpatterns of inheritance typical of other patients with Lynchsyndrome with germlineMMR gene sequence mutations.66

It also appears that cancers tend to occur at a younger agein patients with constitutional MLH1 epimutation (meanage of onset, 39 y) compared with patients with germline

MLH1 sequence mutations (mean age of onset, 44 y).66

MLH1 epimutations are identified by detecting MLH1promoter hypermethylation in DNA derived from normaltissue, such as peripheral blood or non-neoplastic colon. Incases testing positive for an MLH1 epimutation, ideally anadditional source of normal tissue should be tested toconfirm the presence of a constitutional MLH1 epi-mutation. Selection of patients to test for MLH1 epi-mutation is still a subject of debate. One proposedalgorithm is to test patients with anMLH1-deficient canceridentified before the age of 60 years with a personal orfamily history of cancer raising suspicion of Lynch syn-drome and who lack germline sequence mutation inMLH1. The algorithm proposed by some authors does notinclude BRAF, as a subset of patients with constitutionalMLH1 epimutation have tumors that reportedly harborthe BRAF V600E mutation.66

Constitutional MMR DeficiencyConstitutional MMR deficiency (CMMRD) is de-

fined by the presence of biallelic germline mutations inDNA MMR genes. This disease is very rare, with a pro-pensity to present with malignancy early in life, includingearly-onset colorectal carcinoma, lymphoma, leukemia,and brain tumors.72,73 Brain tumors and hematologicalmalignancies are often diagnosed in the first decade of life,whereas intestinal malignancy is more often seen in thesecond and third decades of life.72 Many patients withCMMRD show features reminiscent of neurofibromatosistype 1, such as cafe-au-lait spots.73 In contrast to Lynchsyndrome in which the most common germline mutationsare in MLH1 and MSH2, CMMRD is most often theresult of biallelic mutation of PMS2 (60% of cases) orbiallelic mutation of MSH6 (20% of cases).73 A possiblemechanism for the relatively low frequency of MLH1 orMSH2 mutation in CMMRD is that highly penetrantmutations in MLH1 or MSH2 may be embryologicallylethal when homozygous.74 In contrast to Lynch syn-drome in which loss of expression of the affected MMRgene is only observed in neoplastic tissues, in patients withCMMRD loss of expression of the affected MMR genewill be seen in both neoplastic and non-neoplastic tissues.In young patients with malignancies, lack of MMR ex-pression in non-neoplastic tissues should not necessarilybe interpreted as a failure of the test, provided staining isdemonstrated on external control tissues. In such in-stances, MSI PCR may be helpful to confirm the presenceof MMR protein deficiency within the patient’s tumor.

SCREENING FOR LYNCH SYNDROME INSPECIAL CIRCUMSTANCES

Colorectal AdenomasIn patients suspected of being at risk for Lynch

syndrome and without a personal history of colorectalcarcinoma, evaluation of colorectal adenomas can behelpful in establishing the diagnosis of Lynch syndrome(Fig. 6). A recent series analyzed adenomas in patients withknown deleterious germline mutations in MMR genes




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

and found that loss of MMR protein expression by im-munohistochemistry was identified in 72% (78/109) ofadenomas and correlated with the underlying germlinemutation in all cases.75 Notably, patients with underlyinggermline MSH6 mutations less frequently had adenomasdemonstrating loss of MSH6 protein expression by im-munohistochemistry (4/11, 36%). Adenomas with a villouscomponent or high-grade dysplasia more frequently dem-onstrated loss of MMR protein expression compared withadenomas without a villous component. Other studies havesimilarly identified a significant association between thepresence of high-grade dysplasia and villous componentand loss of MMR protein expression.76 The importance ofadenoma size in selecting polyps to test for MMR proteinimmunohistochemistry is still unresolved. One studydemonstrated that adenoma size does not seem to correlatewith loss of MMR protein expression,75 whereas anotherstudy demonstrated that abnormal MMR protein ex-pression correlated with adenoma size Z8mm.77 Theseresults indicate that assessment of colorectal adenomasusing MMR protein immunohistochemistry can be usefulin patients suspected of having Lynch syndrome, and astrategy of selecting polyps with a villous component, high-grade dysplasia, and/or size Z8mm may increase the yieldof MMR protein immunohistochemistry. Importantly, thepresence of preserved MMR protein expression does notexclude the possibility of Lynch syndrome, as up to 30% to50% of adenomas in patients with known Lynch syndromehave retained MMR protein expression, particularly thosepatients with germline MSH6 mutations.75,77 Sessile ser-rated adenomas can give rise to colorectal carcinoma withsporadic MLH1 deficiency that exhibits MLH1 promoterhypermethylation and the BRAF V600E mutation.78–84

Evaluation of sessile serrated adenomas by MMR proteinimmunohistochemistry has no role in screening for Lynchsyndrome.

The presence of an advanced colorectal adenoma in ayoung patient can raise concern for the possibility of ahereditary cancer syndrome. A recent study analyzedMMR protein immunohistochemistry in advanced ad-enomas identified in patients 45 years and below, definedas adenomas with villous histology, Z1 cm in diameter, orZ3 polyps of any size.85 Only 1 of 66 polyps (1.5%)demonstrated loss of MMR protein expression, indicatingthe low yield of universal MMR protein immunohisto-chemical evaluation of advanced adenomas in youngpatients. A similar analysis using both MMR protein im-munohistochemistry and MSI PCR found that universalscreening of adenomas in patients below 40 years of agedid not identify any patients with abnormal MMR proteinexpression or MSI-H.86 Thus, universal screening of col-orectal adenomas in young patients is not effective inidentifying patients at risk for Lynch syndrome.

Inflammatory Bowel DiseaseInflammatory bowel disease is a well-known risk

factor for colorectal carcinoma, but no guidelines existregarding testing for MMR protein deficiency in this set-ting. In a study of 129 colorectal carcinomas arising ininflammatory bowel disease, MSI-H was observed in 20tumors (15%).87 The pattern of MMR protein expressionby immunohistochemistry was quite variable with loss ofMLH1 and PMS2 expression in 6 colorectal carcinomas,loss of MSH2 and MSH6 expression in 6 colorectal car-cinomas, isolated loss of MSH6 expression in 6 colorectalcarcinomas, and isolated loss of PMS2 expression in 2colorectal carcinomas.MLH1 promoter hypermethylationanalysis performed on the 6 cases with loss of MLH1 andPMS2 expression demonstrated hypermethylation in only3 tumors. Germline mutational analysis was not per-formed. Schulmann et al88 also demonstrated MSI-H in20% of ulcerative colitis–associated colorectal carcinoma

FIGURE 6. In patients suspected of being at risk for Lynch syndrome and without a personal history of colorectal carcinoma,evaluation of colorectal adenomas can be helpful in establishing the diagnosis of Lynch syndrome. In this patient, a colorectaltubular adenoma (A, hematoxylin and eosin) demonstrated loss of MSH6 immunohistochemical expression (B) suggesting Lynchsyndrome. The patient was subsequently found to have a germline deleterious MSH6 mutation.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

and demonstrated that hypermethylation of CpG islandswas the underlying mechanism in most MSI-H ulcerativecolitis–associated colorectal carcinomas. More recently, Liuet al89 performed MMR protein immunohistochemistry on55 inflammatory bowel disease–associated carcinomas anddemonstrated loss of MLH1 and PMS2 expression in4 cancers and loss of MSH2 and MSH6 expression in 1cancer. Liu et al89 also demonstrated that tumors arising inthe setting of inflammatory bowel disease had histologicfeatures commonly associated with MMR protein defi-ciency; however, tumor-infiltrating lymphocytes were lesscommon. These studies demonstrate that MMR proteindeficiency does occur in inflammatory bowel disease-asso-ciated colorectal carcinoma in approximately 10% to 20%of cases. Although germline mutational analysis was notperformed in these studies, it is unlikely that Lynch syn-drome is the underlying cause of colorectal carcinomas inthe setting of inflammatory bowel disease. Rather, mostinflammatory bowel disease–associated colorectal carcino-mas with MMR protein deficiency likely arise because ofhypermethylation of CpG islands in some cases88 andchronic inflammation and reactive oxygen impairing MMRmachinery in others.90 Thus, screening for Lynch syndromein the setting of inflammatory bowel disease–associatedcolorectal carcinoma may not be useful.

Synchronous Intestinal NeoplasiaSynchronous colorectal carcinomas are relatively

rare and account for 1% to 4% of all surgically resectedcarcinomas. MMR protein deficiency is seen in approx-imately 35% of synchronous cancers. Importantly, inmost studies there is a very high concordance of MSIstatus between synchronous tumors, with 87% to 98% ofcases demonstrating either all microsatellite stable orMSI-H tumors. These results suggest common etiologicfactors for neoplasia within a given patient.91,92 Lynchsyndrome patients have a particularly high incidence ofsynchronous carcinomas. Recently, Hu and colleaguesanalyzed 58 patients with 2 synchronous colorectal car-cinomas. In 21 patients, at least 1 tumor was MSI-H, with20 of the 21 tumor pairs demonstrating MSI-H in bothtumors. Of these, 8 (38%) had MMR protein im-munohistochemistry concerning for Lynch syndrome,with the remainder demonstrating BRAF mutations andassociation with sessile serrated adenomas.

Some authors have argued for testing any and allLynch syndrome–associated tumors in a given patient.93

In 1 series, 31% of patients with Lynch syndrome andsynchronous or metachronous LS-associated neoplasiashowed discordant MMR protein expression within in-dividual tumors.93 Given the possibility of discordantMMR protein expression and the potential to miss a di-agnosis of Lynch syndrome if only selected tumors arescreened, it seems prudent to test all Lynch syndrome–associated tumors within a given patient, especially ifthere are clinical features concerning for Lynch syn-drome.

Other Gastrointestinal Tract TumorsPatients with Lynch syndrome are at increased risk

for developing other tumors including within the luminalGI tract, such as adenocarcinomas of the stomach andsmall bowel. Patients with Lynch syndrome have up to300-fold increased risk for developing small bowel ad-enocarcinoma, although the absolute lifetime risk is stillrelatively low at 0.6% to 1%. The frequency of deficientMMR protein expression in small bowel adenocarcinomaranges from 5% to 35%, and Lynch syndrome is over-represented among these tumors compared with color-ectal carcinoma.94 In 1 study of 61 small intestinaladenocarcinomas, 14 were MSI-H and 9 were attributableto Lynch syndrome.95 Thus, it is prudent to screen allsmall bowel adenocarcinomas for Lynch syndrome inpatients with no other apparent risk factors for smallbowel neoplasia.

Although patients with Lynch syndrome are at in-creased risk for developing gastric cancer, the lifetime riskis relatively low at 5%.96 MSI-H occurs in approximately20% of gastric cancers, the vast majority of which arisedue to hypermethylation of the MLH1 promoter. Thesetumors are associated with older patient age, location inthe antrum, well differentiation, and improved survival.Given the relative scarcity of Lynch syndrome–associatedgastric carcinoma, universal screening of gastric carcino-ma by MMR protein immunohistochemistry is generallynot performed. Interestingly, a recent study suggestedthat pyloric gland adenomas may be more common inpatients with Lynch syndrome.97

Current literature suggests that 5% to 10% ofampullary adenocarcinomas exhibit MMR protein defi-ciency.98–100 A large series of 170 ampullary adeno-carcinomas demonstrated MSI-H by PCR in 10% ofampullary adenocarcinomas, with most cases exhibitingintestinal histology associated with tumor-infiltrating lym-phocytes and mucinous differentiation.98 A subset (2%)showed deficient MMR protein expression that raisedconcern for Lynch syndrome. Most of the MSI-H ampul-lary adenocarcinomas demonstrated loss of MLH1 andPMS2 immunohistochemical expression with concomitantMLH1 promoter hypermethylation, indicative of sporadicMLH1 deficiency.98 Agaram and colleagues identifieddeficient MMR protein expression in 3 of 54 cases (6%) ofampullary adenocarcinomas, with 2 cases demonstratingloss of MSH6 expression. Currently, there are no formalguidelines recommending universal screening in ampullaryadenocarcinoma. However, given these data indicating thata small subset of ampullary adenocarcinoma will haveMMR protein deficiency raising concern for Lynch syn-drome, some institutions have opted to perform universalscreening of ampullary adenocarcinoma for Lynch syn-drome with MMR protein immunohistochemistry.

Adenocarcinomas of the appendix infrequently de-monstrate MMR protein deficiency, and most studies havedemonstrated that low-grade mucinous neoplasms of theappendix do not exhibit MMR protein deficiency.101–103

Given the relative scarcity of Lynch syndrome–associatedappendiceal neoplasia, universal screening of appendiceal




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

neoplasia by MMR protein immunohistochemistry is gen-erally not performed.

MANAGEMENT OF PATIENTS WITH LYNCHSYNDROME

Table 5 summarizes the surveillance guidelines forpatients with Lynch syndrome as recommended by the USMulti-Society Task Force for Colorectal Carcinoma.16

The guidelines for screening for colorectal carcinoma bycolonoscopy are strongly recommended. However, al-though the level of evidence supporting the guidelines forendometrial cancer, ovarian cancer, and urinary tractcancer is low, these interventions should be offered topatients at risk for or affected by Lynch syndrome. Thelifetime risk for carcinoma also varies depending on whichMMR gene is affected. Patients with Lynch syndromecaused by MSH6 or PMS2 mutations have a lower cu-mulative lifetime risk for colorectal carcinoma of between10% and 20% compared with patients with Lynch syn-drome caused by MLH1 or MSH2 mutations who have alifetime risk for colorectal carcinoma ranging from 30% to74%.16 In addition, a later onset of colorectal carcinoma isseen in patients with Lynch syndrome caused by MSH6(mean age, 54 to 63 y) or PMS2 (mean age, 47 to 66 y)compared with patients with Lynch syndrome caused byMLH1 or MSH2 mutations (mean age, 27 to 46 y).16

Given this, the initiation of screening for colorectal car-cinoma in patients with Lynch syndrome caused byMSH6 or PMS2 mutation is often delayed compared withLynch syndrome caused by MLH1 or MSH2 mutation.

Patients with Lynch-like syndrome represent a ma-jor challenge in Lynch syndrome management. SomaticMMR gene mutation analysis helps to improve Lynchsyndrome diagnostics by identifying patients with biallelicsomatic MMR gene mutations within their tumor thatlikely do not benefit from the intensive lifelong screening

used for patients with confirmed Lynch syndrome. Usingan updated algorithm for Lynch syndrome diagnostics(Fig. 2), if biallelic somatic mutation of an MMR gene isidentified within the patient’s tumor, the patient does nothave Lynch syndrome and may be screened similarly tothose patients with sporadic carcinoma. However, morestudy is required to determine the risk of developingLynch syndrome–associated carcinoma in patients withLynch-like syndrome and their first-degree relatives toguide screening strategies for these patients.

EMERGING THERANOSTIC IMPLICATIONS OFMMR DEFICIENCY IN GASTROINTESTINAL

TRACT CARCINOMABesides serving as a screening tool for Lynch syn-

drome, testing for MMR protein deficiency is of clinicalimportance as MMR protein deficiency in colorectal car-cinoma is a well-known prognostic and predictive bio-marker. Numerous studies have demonstrated that patientswith MMR protein–deficient colorectal carcinoma have abetter overall survival and reduced recurrence rate com-pared with patients with MMR protein–proficient color-ectal carcinoma.104 This is particularly important forpatients with stage II colorectal carcinoma harboring high-risk histologic features who may be eligible for adjuvantchemotherapy. If the patient has a stage II MMR protein–deficient colorectal carcinoma, adjuvant therapy is gen-erally not given.105,106 In contrast, patients with stage III orIV colorectal carcinoma benefit from chemotherapy re-gardless of MMR protein status within the tumor.107

Immune checkpoint blockade using antibodies tar-geting the programmed cell death protein-1 (PD-1/PD-L1)pathway has emerged as a potential therapeutic option incolorectal carcinoma and other carcinomas of the gas-trointestinal tract. Activation of immune checkpoints al-lows for tumor cells to evade antitumor immunity; PD-1

TABLE 5. Surveillance Guidelines for At-Risk and Affected Patients With Lynch Syndrome*

Interventions Age to Begin IntervalStrength of

Recommendation

Colonoscopy Based on affected MMR gene as follows:MLH1 or MSH2: 20-25 yMSH6: 30 yPMS2: 35 yOr2-5 y younger than youngest age at diagnosis of CRC infamily if diagnosis before the ages listed above

Every 1-2 y; annual colonoscopy should beconsidered in confirmed mutation carriers

Strongrecommenda-tion

Pelvic examination withendometrial sampling

30-35 y Annually Low (expertconsensus)

Transvaginal ultrasound 30-35 y Annually Low (expertconsensus)

EGD with biopsy of theantrum

30-35 y Every 2-3 y based on patient risk factors Low (expertconsensus)

Urinalysis 30-35 y Annually Low (expertconsensus)

*On the basis of guidelines proposed by the US Multi-Society Task Force on Colorectal Carcinoma.16

CRC indicates colorectal carcinoma; EGD, esophagogastroduodenoscopy.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

limits the activity of T cells when it interacts with its ligandPD-L1. Antibodies targeting PD-1 and PD-L1 are beingevaluated in ongoing clinical trials for their efficacy incolorectal carcinoma. MMR protein–deficient colorectalcarcinomas have increased expression of PD-L1 within thetumor and within tumor-infiltrating lymphocytes, allow-ing the tumor to evade antitumor immunity.108 In a recentphase 2 clinical trial, MSI-H status by PCR was predictiveof response to immune checkpoint blockade with pem-brolizumab, a monoclonal antibody that blocks the PD-1pathway.109 The predictive effect of MSI-H and immunecheckpoint blockade was identified not only in colorectalcarcinoma but also in other gastrointestinal tract malig-nancies. Thus, MSI-H has shown early promise as a pre-dictive biomarker for PD-1 blockade immunotherapy.

CONCLUSIONSIdentification of individuals with a hereditary form of

cancer is necessary, as these individuals benefit from geneticcounseling and increased surveillance to prevent cancer de-velopment. By far, Lynch syndrome is the most commonhereditary form of cancer in the gastrointestinal tract. Uni-versal screening of all patients with colorectal carcinoma ei-ther by MMR protein immunohistochemistry or by MSIPCR is now recommended by most major professionalmedical organizations in the United States and Europe. Be-yond the role of tumor diagnosis, pathologists are critical forthe identification of patients at risk for Lynch syndrome andare tasked with establishing screening strategies for patientswith carcinoma of the gastrointestinal tract using ancillarydiagnostic studies, including immunohistochemistry and mo-lecular analyses. In many instances, pathologists are the firstto identify patients at risk for Lynch syndrome and musteffectively communicate these findings to treating clinicians toensure appropriate patient follow-up and genetic evaluation.

REFERENCES1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer

J Clin. 2015;65:5–29.2. Umar A, Boland CR, Terdiman JP, et al. Revised Bethesda

Guidelines for hereditary nonpolyposis colorectal cancer (Lynchsyndrome) and microsatellite instability. J Natl Cancer Inst. 2004;96:261–268.

3. Alexander J, Watanabe T, Wu TT, et al. Histopathologicalidentification of colon cancer with microsatellite instability. Am JPathol. 2001;158:527–535.

4. Risio M, Reato G, di Celle PF, et al. Microsatellite instability isassociated with the histological features of the tumor in nonfamilialcolorectal cancer. Cancer Res. 1996;56:5470–5474.

5. Smyrk TC, Watson P, Kaul K, et al. Tumor-infiltratinglymphocytes are a marker for microsatellite instability in colorectalcarcinoma. Cancer. 2001;91:2417–2422.

6. Shia J, Ellis NA, Paty PB, et al. Value of histopathology inpredicting microsatellite instability in hereditary nonpolyposiscolorectal cancer and sporadic colorectal cancer. Am J SurgPathol. 2003;27:1407–1417.

7. Boland CR, Shike M. Report from the Jerusalem workshop onLynch syndrome-hereditary nonpolyposis colorectal cancer. Gas-troenterology. 2010;138:2197.e1–2197.e7.

8. Moreira L, Balaguer F, Lindor N, et al. Identification of Lynchsyndrome among patients with colorectal cancer. JAMA. 2012;308:1555–1565.

9. Jenkins MA, Hayashi S, O’Shea AM, et al. Pathology features inBethesda guidelines predict colorectal cancer microsatellite insta-bility: a population-based study. Gastroenterology. 2007;133:48–56.

10. Hyde A, Fontaine D, Stuckless S, et al. A histology-based modelfor predicting microsatellite instability in colorectal cancers. Am JSurg Pathol. 2010;34:1820–1829.

11. Hartman DJ, Brand RE, Hu H, et al. Lynch syndrome-associatedcolorectal carcinoma: frequent involvement of the left colon andrectum and late-onset presentation supports a universal screeningapproach. Hum Pathol. 2013;44:2518–2528.

12. Benson AB III, Venook AP, Bekaii-Saab T, et al. Colon cancer,version 3.2014. J Natl Compr Canc Netw. 2014;12:1028–1059.

13. Evaluation of Genomic Applications in Practice, PreventionWorking Group. Recommendations from the EGAPP WorkingGroup: genetic testing strategies in newly diagnosed individualswith colorectal cancer aimed at reducing morbidity and mortalityfrom Lynch syndrome in relatives. Genet Med. 2009;11:35–41.

14. Stoffel EM, Mangu PB, Gruber SB, et al. Hereditary colorectalcancer syndromes: American Society of Clinical Oncology ClinicalPractice Guideline endorsement of the familial risk-colorectalcancer. J Clin Oncol. 2015;33:209–217.

15. Balmana J, Balaguer F, Cervantes A, et al. Familial risk-colorectalcancer. Ann Oncol. 2013;24(suppl 6):vi73–vi80.

16. Giardiello FM, Allen JI, Axilbund JE, et al. Guidelines on geneticevaluation and management of Lynch syndrome: a consensusstatement by the US Multi-Society Task Force on colorectalcancer. Gastroenterology. 2014;147:502–526.

17. Syngal S, Brand RE, Church JM, et al. ACG clinical guideline:Genetic testing and management of hereditary gastrointestinalcancer syndromes. Am J Gastroenterol. 2015;110:223–262. Quiz 63.

18. Mvundura M, Grosse SD, Hampel H, et al. The cost-effectivenessof genetic testing strategies for Lynch syndrome among newlydiagnosed patients with colorectal cancer. Genet Med. 2010;12:93–104.

19. de Jong AE, van Puijenbroek M, Hendriks Y, et al. Microsatelliteinstability, immunohistochemistry, and additional PMS2 stainingin suspected hereditary nonpolyposis colorectal cancer. Clin CancerRes. 2004;10:972–980.

20. Southey MC, Jenkins MA, Mead L, et al. Use of molecular tumorcharacteristics to prioritize mismatch repair gene testing in early-onset colorectal cancer. J Clin Oncol. 2005;23:6524–6532.

21. Hampel H, Frankel WL, Martin E, et al. Screening for the Lynchsyndrome (hereditary nonpolyposis colorectal cancer). N Engl JMed. 2005;352:1851–1860.

22. Mojtahed A, Schrijver I, Ford JM, et al. A two-antibody mismatchrepair protein immunohistochemistry screening approach forcolorectal carcinomas, skin sebaceous tumors, and gynecologictract carcinomas. Mod Pathol. 2011;24:1004–1014.

23. Shia J, Tang LH, Vakiani E, et al. Immunohistochemistry as first-line screening for detecting colorectal cancer patients at risk forhereditary nonpolyposis colorectal cancer syndrome: a 2-antibodypanel may be as predictive as a 4-antibody panel. Am J SurgPathol. 2009;33:1639–1645.

24. Hall G, Clarkson A, Shi A, et al. Immunohistochemistry for PMS2and MSH6 alone can replace a four antibody panel for mismatchrepair deficiency screening in colorectal adenocarcinoma. Pathol-ogy. 2010;42:409–413.

25. Overbeek LI, Ligtenberg MJ, Willems RW, et al. Interpretationof immunohistochemistry for mismatch repair proteins is onlyreliable in a specialized setting. Am J Surg Pathol. 2008;32:1246–1251.

26. Radu OM, Nikiforova MN, Farkas LM, et al. Challenging casesencountered in colorectal cancer screening for Lynch syndromereveal novel findings: nucleolar MSH6 staining and impact of priorchemoradiation therapy. Hum Pathol. 2011;42:1247–1258.

27. Mihaylova VT, Bindra RS, Yuan J, et al. Decreased expression ofthe DNA mismatch repair gene Mlh1 under hypoxic stress inmammalian cells. Mol Cell Biol. 2003;23:3265–3273.

28. Chang CL, Marra G, Chauhan DP, et al. Oxidative stressinactivates the human DNA mismatch repair system. Am J PhysiolCell Physiol. 2002;283:C148–C154.




Elaine Huang

Elaine Huang

Elaine Huang

Elaine Huang

29. Bao F, Panarelli NC, Rennert H, et al. Neoadjuvant therapyinduces loss of MSH6 expression in colorectal carcinoma. Am JSurg Pathol. 2010;34:1798–1804.

30. Vilkin A, Halpern M, Morgenstern S, et al. How reliable isimmunohistochemical staining for DNA mismatch repair proteinsperformed after neoadjuvant chemoradiation? Hum Pathol.2014;45:2029–2036.

31. Joost P, Veurink N, Holck S, et al. Heterogenous mismatch-repairstatus in colorectal cancer. Diagn Pathol. 2014;9:126–136.

32. Pai RK, Plesec TP, Abdul-Karim FW, et al. Abrupt loss of MLH1and PMS2 expression in endometrial carcinoma: molecular andmorphologic analysis of 6 cases. Am J Surg Pathol. 2015;39:993–999.

33. Shia J, Zhang L, Shike M, et al. Secondary mutation in a codingmononucleotide tract in MSH6 causes loss of immunoexpression ofMSH6 in colorectal carcinomas with MLH1/PMS2 deficiency.Mod Pathol. 2013;26:131–138.

34. Raymond VM, Morris AM, Hafez KS, et al. MLH1 promotorhypermethylation does not rule out a diagnosis of Lynchsyndrome: a case report. Fam Cancer. 2015;14:77–80.

35. Dudley B, Brand RE, Thull D, et al. Germline MLH1 mutationsare frequently identified in Lynch syndrome patients with color-ectal and endometrial carcinoma demonstrating isolated loss ofPMS2 immunohistochemical expression. Am J Surg Pathol. 2015;39:1114–1120.

36. Suraweera N, Duval A, Reperant M, et al. Evaluation of tumormicrosatellite instability using five quasimonomorphic mononu-cleotide repeats and pentaplex PCR. Gastroenterology. 2002;123:1804–1811.

37. Funkhouser WK Jr, Lubin IM, Monzon FA, et al. Relevance,pathogenesis, and testing algorithm for mismatch repair-defectivecolorectal carcinomas: a report of the association for molecularpathology. J Mol Diagn. 2012;14:91–103.

38. Vilar E, Mork ME, Cuddy A, et al. Role of microsatelliteinstability-low as a diagnostic biomarker of Lynch syndrome incolorectal cancer. Cancer Genet. 2014;207:495–502.

39. You JF, Buhard O, Ligtenberg MJ, et al. Tumours with loss ofMSH6 expression are MSI-H when screened with a pentaplex offive mononucleotide repeats. Br J Cancer. 2010;103:1840–1845.

40. Salipante SJ, Scroggins SM, Hampel HL, et al. Microsatelliteinstability detection by next generation sequencing. Clin Chem.2014;60:1192–1199.

41. Moreira L, Munoz J, Cuatrecasas M, et al. Prevalence of somaticmutl homolog 1 promoter hypermethylation in Lynch syndromecolorectal cancer. Cancer. 2015;121:1395–1404.

42. McGivern A, Wynter CV, Whitehall VL, et al. Promoter hyper-methylation frequency and BRAF mutations distinguish hereditarynon-polyposis colon cancer from sporadic MSI-H colon cancer.Fam Cancer. 2004;3:101–107.

43. Jin M, Hampel H, Zhou X, et al. BRAF V600E mutation analysissimplifies the testing algorithm for Lynch syndrome. Am J ClinPathol. 2013;140:177–183.

44. Parsons MT, Buchanan DD, Thompson B, et al. Correlation oftumour BRAF mutations and MLH1 methylation with germlinemismatch repair (MMR) gene mutation status: a literature reviewassessing utility of tumour features for MMR variant classification.J Med Genet. 2012;49:151–157.

45. Roth RM, Hampel H, Arnold CA, et al. A modified Lynchsyndrome screening algorithm in colon cancer: BRAF immuno-histochemistry is efficacious and cost beneficial. Am J Clin Pathol.2015;143:336–343.

46. Kuan SF, Navina S, Cressman KL, et al. Immunohistochemicaldetection of BRAF V600E mutant protein using the VE1 antibodyin colorectal carcinoma is highly concordant with molecular testingbut requires rigorous antibody optimization. Hum Pathol. 2014;45:464–472.

47. Adackapara CA, Sholl LM, Barletta JA, et al. Immunohistochem-istry using the BRAF V600E mutation-specific monoclonal antibodyVE1 is not a useful surrogate for genotyping in colorectaladenocarcinoma. Histopathology. 2013;63:187–193.

48. Heald B, Plesec T, Liu X, et al. Implementation of universalmicrosatellite instability and immunohistochemistry screeningfor diagnosing lynch syndrome in a large academic medical center.J Clin Oncol. 2013;31:1336–1340.

49. Peltomaki P, Offerhaus GJA, Vasen H. Lynch syndrome. In:Bosman FT, Carneiro F, Hruban RH, Theise ND, eds. WHOClassification of Tumours of the Digestive System. Lyon, France:IARC Press; 2010:152–155.

50. Lindor NM, Rabe K, Petersen GM, et al. Lower cancer incidencein Amsterdam-I criteria families without mismatch repair defi-ciency: familial colorectal cancer type X. JAMA. 2005;293:1979–1985.

51. Shiovitz S, Copeland WK, Passarelli MN, et al. Characterisation offamilial colorectal cancer Type X, Lynch syndrome, and non-familial colorectal cancer. Br J Cancer. 2014;111:598–602.

52. Dominguez-Valentin M, Therkildsen C, Da Silva S, et al. Familialcolorectal cancer type X: genetic profiles and phenotypic features.Mod Pathol. 2015;28:30–36.

53. Nieminen TT, O’Donohue MF, Wu Y, et al. Germline mutation ofRPS20, encoding a ribosomal protein, causes predisposition tohereditary nonpolyposis colorectal carcinoma without DNA mis-match repair deficiency. Gastroenterology. 2014;147:595–598. e5.

54. Schulz E, Klampfl P, Holzapfel S, et al. Germline variants in theSEMA4A gene predispose to familial colorectal cancer type X. NatCommun. 2014;5:5191.

55. Wei C, Peng B, Han Y, et al. Mutations of HNRNPA0 and WIF1predispose members of a large family to multiple cancers. FamCancer. 2015;14:297–306.

56. Rodriguez-Soler M, Perez-Carbonell L, Guarinos C, et al. Risk ofcancer in cases of suspected Lynch syndrome without germlinemutation. Gastroenterology. 2013;144:926–932.

57. Win AK, Buchanan DD, Rosty C, et al. Role of tumour molecularand pathology features to estimate colorectal cancer risk for first-degree relatives. Gut. 2015;64:101–110.

58. Buchanan DD, Rosty C, Clendenning M, et al. Clinical problemsof colorectal cancer and endometrial cancer cases with unknowncause of tumor mismatch repair deficiency (suspected Lynchsyndrome). Appl Clin Genet. 2014;7:183–193.

59. Mas-Moya J, Dudley B, Brand RE, et al. Clinicopathologicalcomparison of colorectal and endometrial carcinomas in patientswith Lynch-like syndrome versus patients with Lynch syndrome.Hum Pathol. 2015;46:1616–1625.

60. Rhees J, Arnold M, Boland CR. Inversion of exons 1-7 of theMSH2 gene is a frequent cause of unexplained Lynch syndrome inone local population. Fam Cancer. 2014;13:219–225.

61. Sourrouille I, Coulet F, Lefevre JH, et al. Somatic mosaicism anddouble somatic hits can lead to MSI colorectal tumors. FamCancer. 2013;12:27–33.

62. Castillejo A, Vargas G, Castillejo MI, et al. Prevalence of germlineMUTYH mutations among Lynch-like syndrome patients. Eur JCancer. 2014;50:2241–2250.

63. Mensenkamp AR, Vogelaar IP, van Zelst-Stams WA, et al.Somatic mutations in MLH1 and MSH2 are a frequent cause ofmismatch-repair deficiency in Lynch syndrome-like tumors.Gastroenterology. 2014;146:643–646. e8.

64. Geurts-Giele WR, Leenen CH, Dubbink HJ, et al. Somaticaberrations of mismatch repair genes as a cause of microsatellite-unstable cancers. J Pathol. 2014;234:548–559.

65. Haraldsdottir S, Hampel H, Tomsic J, et al. Colon and endometrialcancers with mismatch repair deficiency can arise from somatic, ratherthan germline, mutations. Gastroenterology. 2014;147:1308–1316. e1.

66. Hitchins MP. The role of epigenetics in Lynch syndrome. FamCancer. 2013;12:189–205.

67. Ligtenberg MJ, Kuiper RP, Geurts van Kessel A, et al. EPCAMdeletion carriers constitute a unique subgroup of Lynch syndromepatients. Fam Cancer. 2013;12:169–174.

68. Rumilla K, Schowalter KV, Lindor NM, et al. Frequency ofdeletions of EPCAM (TACSTD1) in MSH2-associated Lynchsyndrome cases. J Mol Diagn. 2011;13:93–99.




69. Kloor M, Voigt AY, Schackert HK, et al. Analysis of EPCAMprotein expression in diagnostics of Lynch syndrome. J Clin Oncol.2011;29:223–227.

70. Huth C, Kloor M, Voigt AY, et al. The molecular basis of EPCAMexpression loss in Lynch syndrome-associated tumors.Mod Pathol.2012;25:911–916.

71. Peltomaki P. Epigenetic mechanisms in the pathogenesis of Lynchsyndrome. Clin Genet. 2014;85:403–412.

72. Vasen HF, Ghorbanoghli Z, Bourdeaut F, et al. Guidelines forsurveillance of individuals with constitutional mismatch repair-deficiency proposed by the European Consortium “Care forCMMR-D” (C4CMMR-D). J Med Genet. 2014;51:283–293.

73. Wimmer K, Kratz CP, Vasen HF, et al. Diagnostic criteria forconstitutional mismatch repair deficiency syndrome: suggestions ofthe European consortium’care for CMMRD’ (C4CMMRD).JJ Med Genet. 2014;51:355–365.

74. Bakry D, Aronson M, Durno C, et al. Genetic and clinicaldeterminants of constitutional mismatch repair deficiency syn-drome: report from the constitutional mismatch repair deficiencyconsortium. Eur J Cancer. 2014;50:987–996.

75. Walsh MD, Buchanan DD, Pearson SA, et al. Immunohistochem-ical testing of conventional adenomas for loss of expression ofmismatch repair proteins in Lynch syndrome mutation carriers: acase series from the Australasian site of the colon cancer familyregistry. Mod Pathol. 2012;25:722–730.

76. Pino MS, Mino-Kenudson M, Wildemore BM, et al. DeficientDNA mismatch repair is common in Lynch syndrome-associatedcolorectal adenomas. J Mol Diagn. 2009;11:238–247.

77. Yurgelun MB, Goel A, Hornick JL, et al. Microsatellite instabilityand DNA mismatch repair protein deficiency in Lynch syndromecolorectal polyps. Cancer Prev Res (Phila). 2012;5:574–582.

78. Herman JG, Umar A, Polyak K, et al. Incidence and functionalconsequences of hMLH1 promoter hypermethylation in colorectalcarcinoma. Proc Natl Acad Sci. 1998;95:6870–6875.

79. Kane MF, Loda M, Gaida GM, et al. Methylation of the hMLH1promoter correlates with lack of expression of hMLH1 in sporadiccolon tumors and mismatch repair-defective human tumor celllines. Cancer Res. 1997;57:808–811.

80. Kambara T, Simms LA, Whitehall VL, et al. BRAF mutation isassociated with DNAmethylation in serrated polyps and cancers ofthe colorectum. Gut. 2004;53:1137–1144.

81. Carr NJ, Mahajan H, Tan KL, et al. Serrated and non-serratedpolyps of the colorectum: their prevalence in an unselected case seriesand correlation of BRAF mutation analysis with the diagnosis ofsessile serrated adenoma. J Clin Pathol. 2009;62:516–518.

82. Weisenberger DJ, Siegmund KD, Campan M, et al. CpG islandmethylator phenotype underlies sporadic microsatellite instabilityand is tightly associated with BRAF mutation in colorectal cancer.Nat Genet. 2006;38:787–793.

83. O’Brien MJ, Yang S, Mack C, et al. Comparison of microsatelliteinstability, CpG island methylation phenotype, BRAF and KRASstatus in serrated polyps and traditional adenomas indicatesseparate pathways to distinct colorectal carcinoma end points.Am J Surg Pathol. 2006;30:1491–1501.

84. Spring KJ, Zhao ZZ, Karamatic R, et al. High prevalence of sessileserrated adenomas with BRAF mutations: a prospective study ofpatients undergoing colonoscopy. Gastroenterology. 2006;131:1400–1407.

85. Kushnir VM, Nalbantoglu IL, Watson R, et al. Advancedcolorectal adenomas in patients under 45 years of age are mostlysporadic. Dig Dis Sci. 2014;59:2757–2764.

86. Velayos FS, Allen BA, Conrad PG, et al. Low rate of microsatelliteinstability in young patients with adenomas: reassessing theBethesda guidelines. Am J Gastroenterol. 2005;100:1143–1149.

87. Svrcek M, El-Bchiri J, Chalastanis A, et al. Specific clinical andbiological features characterize inflammatory bowel disease asso-ciated colorectal cancers showing microsatellite instability. J ClinOncol. 2007;25:4231–4238.

88. Schulmann K, Mori Y, Croog V, et al. Molecular phenotype ofinflammatory bowel disease-associated neoplasms with micro-satellite instability. Gastroenterology. 2005;129:74–85.

89. Liu X, Goldblum JR, Zhao Z, et al. Distinct clinicohistologicfeatures of inflammatory bowel disease-associated colorectaladenocarcinoma: in comparison with sporadic microsatellite-stableand Lynch syndrome-related colorectal adenocarcinoma. Am JSurg Pathol. 2012;36:1228–1233.

90. Hofseth LJ, Khan MA, Ambrose M, et al. The adaptive imbalancein base excision-repair enzymes generates microsatellite instabilityin chronic inflammation. J Clin Invest. 2003;112:1887–1894.

91. Dykes SL, Qui H, Rothenberger DA, et al. Evidence of a preferredmolecular pathway in patients with synchronous colorectal cancer.Cancer. 2003;98:48–54.

92. Hu H, Chang DT, Nikiforova MN, et al. Clinicopathologicfeatures of synchronous colorectal carcinoma: A distinct subsetarising from multiple sessile serrated adenomas and associated withhigh levels of microsatellite instability and favorable prognosis. AmJ Surg Pathol. 2013;37:1660–1670.

93. Roth R, Hampel H, Arnold CA, et al. Discordant mismatch repairprotein immunoreactivity in Lynch syndrome-associated neo-plasms: a recommendation for screening synchronous/metachro-nous lesions. Mod Pathol. 2015;28(suppl 2):189A.

94. Aparicio T, Zaanan A, Svrcek M, et al. Small bowel adenocarci-noma: epidemiology, risk factors, diagnosis and treatment. DigLiver Dis. 2014;46:97–104.

95. Aparicio T, Svrcek M, Zaanan A, et al. Small bowel adenocarci-noma phenotyping, a clinicobiological prognostic study. Br JCancer. 2013;109:3057–3066.

96. Capelle LG, Van Grieken NC, Lingsma HF, et al. Risk andepidemiological time trends of gastric cancer in Lynch syndromecarriers in the Netherlands. Gastroenterology. 2010;138:487–492.

97. Lee SE, Kang SY, Cho J, et al. Pyloric gland adenoma in Lynchsyndrome. Am J Surg Pathol. 2014;38:784–792.

98. Ruemmele P, Dietmaier W, Terracciano L, et al. Histopathologicfeatures and microsatellite instability of cancers of the papilla of vaterand their precursor lesions. Am J Surg Pathol. 2009;33:691–704.

99. Agaram NP, Shia J, Tang LH, et al. DNA mismatch repairdeficiency in ampullary carcinoma: a morphologic and immunohis-tochemical study of 54 cases. Am J Clin Pathol. 2010;133:772–780.

100. Sessa F, Furlan D, Zampatti C, et al. Prognostic factors forampullary adenocarcinomas: tumor stage, tumor histology, tumorlocation, immunohistochemistry and microsatellite instability.Virchows Archiv. 2007;451:649–657.

101. Taggart MW, Galbincea J, Mansfield PF, et al. High-levelmicrosatellite instability in appendiceal carcinomas. Am J SurgPathol. 2013;37:1192–1200.

102. Misdraji J, Burgart LJ, Lauwers GY. Defective mismatch repair inthe pathogenesis of low-grade appendiceal mucinous neoplasmsand adenocarcinomas. Mod Pathol. 2004;17:1447–1454.

103. Zauber P, Berman E, Marotta S, et al. Ki-ras gene mutations areinvariably present in low-grade mucinous tumors of the vermiformappendix. Scand J Gastroenterol. 2011;46:869–874.

104. Hutchins G, Southward K, Handley K, et al. Value of mismatchrepair, KRAS, and BRAF mutations in predicting recurrence andbenefits from chemotherapy in colorectal cancer. J Clin Oncol.2011;29:1261–1270.

105. Sargent DJ, Marsoni S, Monges G, et al. Defective mismatchrepair as a predictive marker for lack of efficacy of fluorouracil-based adjuvant therapy in colon cancer. J Clin Oncol. 2010;28:3219–3226.

106. Dienstmann R, Salazar R, Tabernero J. Personalizing colon canceradjuvant therapy: selecting optimal treatments for individualpatients. J Clin Oncol. 2015;33:1787–1796.

107. Gavin PG, Colangelo LH, Fumagalli D, et al. Mutation profilingand microsatellite instability in stage II and III colon cancer: anassessment of their prognostic and oxaliplatin predictive value. ClinCancer Res. 2012;18:6531–6541.

108. Llosa NJ, Cruise M, Tam A, et al. The vigorous immunemicroenvironment of microsatellite instable colon cancer is bal-anced by multiple counter-inhibitory checkpoints. Cancer Discov.2015;5:43–51.

109. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors withmismatch-repair deficiency. N Engl J Med. 2015;372:2509–2520.




Elaine Huang

Elaine Huang

chi meisub.chimei.org.tw/55399/images/pdf/04_announcement/2017/ap-an-106-07.pdf · rish k. pai, md,...

Documents